Abstract
It is well known that the expected waiting time for customers routed to several parallel queues decreases dramatically when customers are routed to the shortest of two randomly chosen queues, rather than being arbitrarily assigned to one of the queues, and that the further improvement when there are three queues to choose from is much less than the improvement when moving from one to two queues (the power of two [5]). We consider the power of two effect when a subset of customers are flexible, and can choose the shortest of two queues, while the remainder are dedicated, and have no routing choice. We show that the stationary expected waiting time is decreasing and convex in the proportion of flexible customers. Our results show that having a small proportion of flexible customers has nearly as much benefit as having full power of two choices.
- O. T. Akgun, R. Righter, and R. Wolff. Multiple server system with flexible arrivals. Adv. Appl. Probab., to appear.Google Scholar
- O. T. Akgun, R. Righter, and R. Wolff. Understanding the marginal impact of customer flexibility. Queueing Syst., to appear. Google ScholarDigital Library
- D. J. Daley and T. Rolski. Finiteness of waiting-time moments in general stationary single server queues. Ann. Appl. Probab., 2:987--1008, 1992.Google ScholarCross Ref
- Y. T. He and D. Down. On accommodating customer flexibility in service systems. INFOR., 47:289--295, 2009.Google Scholar
- M. Mitzenmacher. The power of two choices in randomized load balancing. IEEE Trans. Parallel and Distrib. Syst., 12:1094--1104, 2001. Google ScholarDigital Library
Index Terms
The power of partial power of two choices
Recommendations
Understanding the marginal impact of customer flexibility
We study the marginal impact of customer flexibility in service systems. We consider a queueing system with multiple parallel servers, in which a proportion of customers are flexible and can go to any server, while the remainder require service at a ...
Insensitivity for PS server farms with JSQ routing
Join-the-Shortest-Queue (JSQ) is a very old and popular routing policy for server farms. Figure 1 shows two examples of server farm architectures employing JSQ routing. In both cases, each incoming job is immediately dispatched, via a front-end router, ...
The Power of Two Choices in Randomized Load Balancing
We consider the following natural model: Customers arrive as a Poisson stream of rate \lambda n, \lambda < 1, at a collection of n servers. Each customer chooses some constant d servers independently and uniformly at random from the n servers and waits ...
Comments